October - 2000 - issue > Cover

On August 23, more than 51 million TV viewers watched as a rather arrogant and previously unknown corporate trainer named Richard Hatch enjoyed a fleeting moment as the world’s most famous human being -- crowned the million-dollar winner of ABC’s absurd but popular reality show, “Survivor.” Quick to capitalize on the success of the series before the world comes to its senses, CBS has already slated the next set of episodes for the Australian outback. But rather than place the next crew of “castaways” among the kangaroos, the CBS team might consider an alternate location--Silicon Valley.

In all seriousness, the Valley entrepreneurial experience is about living on the brink of elimination, and although the rise from a few guys and a good idea to a working business is never just a smooth climb to success and riches, the payoffs in the Valley game can also be more than just a million bucks and a new car. Witness Lara Networks, a company with a story of perseverance and unexpected twists and turns that might be worthy of its own TV series. This is “Survivor, Valley style” - only it also happens to be a story much more worthwhile for the rest of the world.

In 1996, Jayan Ramankutty, Ajit Medhekar and Kamal Gunsagar -- all former colleagues at Alliance Semiconductor -- got together to build a company. They brought extensive domain experience within the many facets of the semiconductor business, as well as a substantial network of industry contacts. They understood startups and their experience with Alliance, which went public in 1993, had been a rewarding one. They also had what they were convinced was the all-important good idea.

The idea, according to co-founder Ajit Medhekar, was based on the fact that “the networking industry was the next industry about to be ‘siliconized.’” He explains, “The technology [in the networking industry] was based on software. Companies were building from off-the-shelf semiconductor components and then putting great software in.” It was clear to the founders of Lara that this technology would only go so far, so they set about conceptualizing a product that could more effectively do the all-important “decision making” for IP packets traveling across the network.

The answer was in what is termed a “search engine” -- nothing to do with the algorithms and colorful links of Yahoo or Google, but rather a chip that instantaneously determines where an IP packet is coming from and where it needs to go. The core technology, CAM (content addressable memories), was one that had been around in the memory world for a long time without ever finding its real killer application. CAM technology would allow the Lara search engine to take information about an IP packet and compare it, instantaneously and simultaneously, with stored information so as to immediately understand what the packet is and where it needs to go.

This product would, in the minds of the Lara team, allow networks to run much more efficiently and reliably. “All of it made sense on paper,” explains Medhekar. But rather than build the product and then try to sell it, the strategy would be, in the words of Jayan Ramankutty, to “sell the idea first, get a buy-in from a customer, and build the product per the customer’s requirements.”

The team took a road trip from coast to coast, speaking to people like Desh Deshpande (at that time with Cascade) and Yago founder Piyush Patel. Although most showed enthusiasm for the idea, none would truly endorse the product. And yet it was with Cisco that the Lara team placed their most profound hopes. “I met with Cisco six or seven times,” explains Ramankutty, “late in the evening, early in the morning -- it didn’t matter.” But the team still wasn’t able to nail down any outside commitment to the product.

What made matters more urgent was the fact that the three men were financing the operation themselves. Says Raman-kutty, “We had a 4,000-square foot facility, we had four engineers on board designing the product, we had bought workstations.” At this point, the Lara team was out a lot of money with not a lot to show for it, and they were beginning to question whether this was such a hot idea. To make matters worse, on June 7, a major semiconductor company committed an important $4 million to Lara, only to back out on the following day. By September, according to Ramankutty, “It was a pretty bad scene.”

In October of ’97, on the verge of abandoning their initial business model, the team managed to garner the attention of Cisco’s gigabit group, headed by Valley superstar Andy Bechtolsheim. Bechtolsheim saw real merit in what the Lara team was trying to achieve, but due to various conflicts of interest within Cisco, it was not until several meetings later that the networking giant finally provided the buy-in that was so important to the three founders.

Says Ramankutty, “Having Cisco as a customer was the best possible outcome.” In the nick of time, the struggling company had obtained the customer it had been seeking from the start; Lara delivered its first product to Cisco in June of ’98, one year after the real start of the business. The company raised a venture round of $2 million, and an additional $2 million from private sources.

But the road ahead would prove to be much longer and complicated than the Lara founders had ever anticipated.

One Company, Two Faces

When the Lara team met one man, Raj Saksena, who was working at Nortel on voice and data convergence within unified networks, everything changed. Saksena was convinced that the Lara search engine could be put to work in the voice-over IP (VoIP) market. Soon, with the team sold on Saksena’s idea, Lara set out to build a convergence processor.

Unlike the earlier Lara chip, this media processor was quick to gain attention, from people in the industry hoping to buy it as well as from eager investors seeking a hot opportunity to grab a piece of the pie. But the three men, now backed up by a small core team, once again chose to take an unconventional approach. Medhekar explains: “We sat back and said, ‘this is too good a chip to just go out and sell as a chip -- let’s build a system around it.’” He hastens to add, “In this business, going from silicon to systems is a huge leap of faith.”

It was June of 1999, and the Lara investors were taken aback, alarmed that their investment in a chips company had suddenly become an investment in a company doing network components and voice-over IP systems, all under one roof. Medhekar recalls his promise: “I will have a signed customer by September.” Medhekar got the industry vote of confidence that he needed and soon the VoIP division of the company was hurtling into the market as an autonomous division of Lara.

But now the company had an identity crisis on its hands. The Interop show in 1999 provided a climax. Says Jayan Ramankutty, “We had a booth where we had this big system on one side and a small part on the other side with a bunch of modules -- we didn’t know what was going on. We were trying to get two businesses going at once.” On top of that, some Lara customers were becoming potential competitors of the new VoIP system.

In December 1999, Lara Networks spun off as a 24-person team, and the VoIP systems end adopted a new name -- Empowertel. The two companies have beco-me totally independent, though Raman-kutty admits that there is “some level of agreement” between the two to ensure optimum profitability on both ends. In June of this year, Ramankutty moved permanently to Lara as president and COO, with Gunsagar as CEO. Medhekar has become the president and CEO of Empowertel.

And so the three former colleagues, who so struggled to get Lara off the ground, have seen their unpredictable journey produce two completely independent companies. Says Medhekar, “It’s been a fun ride.”

The Future

Lara Networks has healthy revenues and can hardly meet demand for their search engines. “Our main objective is to grow Lara into a full-fledged silicon provider for the whole communication and telecom space,” says Jayan Ramankutty. For now, the company seems to be on the right track, despite the loss of a significant amount of momentum as a result of the Empowertel split. Still not content to just sit on a good idea, the company is looking to a future that will include not just search engine chips, but also other kinds of networking processors and even broadband switches, all of which will be targeted at next-generation optical networks. Now numbering over 80 people, Lara is currently completing a $40 million series B round of funding, and the company is moving rapidly down the path toward a future public offering. The goal, according to Ramankutty, is to “very quickly create a class-A company of a caliber similar to PMC, Broadcom or AMCC.”

For Empowertel, will the systems gamble pay off?

Raj Saksena, the man behind the idea, is on board as CTO and vice president of business development. “This is truly a breakthrough technology,” says Medhekar. “My belief is that it’s going to change the way the world communicates. This is what communication is about. It’s not browsing on the Internet -- the real thing is voice.” Medhekar explains that the hype surrounding VoIP hasn’t produced any real, or good quality, solutions. As a result, the size of the IP telephony market has failed miserably to live up to market projections. Empowertel hopes to fill the void.

“Once I’ve got voice on an IP network, what can I do with it? That’s the story we tell,” Medhekar explains enthusiastically. Empowertel solutions will seek to provide email, text-to speech, voice mail and other services - with the ability to engage all of them, seamlessly, during a call. The concept is “less voice-over IP,” Medhekar explains, than it is “voice and IP.” The company now numbers over 200 people worldwide, has just raised $54 million dollars. Empowertel hopes to be shipping products before the year is out.